Issue link: https://iconnect007.uberflip.com/i/948150
60 SMT007 MAGAZINE I MARCH 2018 based system. Other than the simple user inter- face, none of the software is on your device, and other than the result of the search, none of the data. Search engines are designed to be extremely well optimized to cope with, let's call it "human" data, the majority of which was not created with searching in mind. It then takes a human mind to select the results that make sense, which may not be the one at the top of the list. As the internet has matured, search engine optimization (SEO) techniques have been created which embed key items of data within web pages designed to help search engines understand the content and intent of the web data. Even for humans, just random data in the cloud is not good enough. Considerations of Cloud Storage Web pages on the Internet do not of course reflect the true nature of cloud data storage. From a typical manufacturing standpoint, cloud storage is simply another way or place in which to put data. From a usage perspec- tive, many see it as being no different from an on-site server (now marketed as a "local cloud") or even the external hard disk on your laptop (now marketed as your "personal cloud"). Off-site cloud storage is vast and can easily be cost effective, requiring no mainte- nance or fixed overhead other than paying the service bills. There are however a couple of very important things to consider with cloud storage. Firstly, you must trust that there is appro- priate security in place, and in some cases, for example with ITAR restricted data, make sure that the physical storage distribution is within friendly places. Cloud data is generally phys- ically spread across data-centers all over the world and will naturally be driven towards areas that offer lower costs. Adequate security for data access must be provided. The second thing to consider with cloud data storage is the requirement for how the data will be accessed. Links from a site to the cloud have a finite speed, which is basi- cally the Internet connection. Though theoret- ical upload and download speeds may seem quite fast these days, these same connections are supporting the email system, as well as everyone in the company on a browser, and even the coffee machine these days. Mission critical systems must share bandwidth with an increasing number of devices and services unless equipped with dedicated lines. Limited connection speed is all very well for the continuous trickle of upload data, but once up there, it is rather impractical to ever consider downloading all the data again. As with the search engine example, if your data and software are both in the cloud, system performance and capability will not be limited by connection issues. In many use-cases, this is fine, for example the use of enterprise-grade Business Intelligence tools to look at longer term statistical trends. The Practical Bottleneck With a general search, like standard Google in a browser, we as humans have to try to inter- pret and select the results that make sense. We see now however that technology such as Siri, Cortana, Alexa, Echo, etc., are trying to become much smarter, to be able to deliver an answer to a question which is not just a list of web pages that might be of interest. To do this, there is another level of data organization required, over and above standard SEO techniques. It has taken quite some time, and a huge amount of investment to make this happen by companies behind these products, as they are trying to convert random data into meaningful informa- tion. This has always been the same key bottle- neck for any computer system, cloud-based or Off-site cloud storage is vast and can easily be cost effective, requiring no maintenance or fixed overhead other than paying the service bills.